blues from music · microsoft powerpoint - blues from music.ppt author: str_msp created date:...

BLUES from Music:BLind Underdetermined Extraction of Sources

from Music

Michael Syskind Pedersen Tue Lehn-Schiøler

Jan LarsenIMM, Technical University of Denmark

ICA2006, Charleston, SC, USA

Separating music into basic components

Michael Syskind Pedersen, IMM, Technical University of Denmark

Motivation: Why separating music?

• Music Transcription• Identifying instruments• Identify vocalist


Assumptions• Stereo recording of the music piece is

available.• The instruments are separated to some

extent in time and in frequency, i.e. the instruments are sparse in the time-frequency (T-F) domain.

• The different instruments originate from spatially different directions.


Separation principle 1: T-F masking

Stereo channel 1 Stereo channel 2

Gain difference between channels

Separation principle 2: ICA

sources mixedsignals

recovered source signals

mixing

x = As

separation

ICAy = Wx

What happens if a 2-by-2 separation matrix W is applied

to a 2-by-N mixing system?


ICA on stereo signals

• We assume that the mixture can be modeled as an instantaneous mixture, i.e.

• The ratio between the gains in each column in the mixing matrix corresponds to a certain direction.

⎥⎦

⎤⎢⎣

⎡=

)()()()(

)(212

111

N

N

rrrr

Aθθθθ

θL

LsAx N ), ... ,( 1 θθ=


Direction dependent gain|)(|log20)( θWAθr =

When W is applied, the two separated channels each contain a group of sources, which is as independent as possible from the other channel.


x1 x2

ICA

STFT STFT

y1 y2

Y1(t, f) Y2(t, f)

⎩⎨⎧ >

= otherwise 0

when 1

c / YYBM 21

1⎩⎨⎧ >

= otherwise 0

when 1

c / YYBM 12

2

X1(t,f)

BM1 BM2

x1(2) x2

(2)

ICA+BM Separator

^ ^

Combining ICA and T-F masking

ISTFT

X2(t,f)

ISTFT

X1(t,f)

x1(2) x2

(2)^ ^ISTFT

X2(t,f)

ISTFT

Method applied iterativelyx1 x2

ICA+BM

ICA+BM ICA+BM

ICA+BM ICA+BM


Improving method• The assumption of

instantaneous mixing may not always hold.

• Assumption can be relaxed.

• Separation procedure is continued until very sparse masks are obtained.

• Masks that mainly contain the same source are afterwards merged.

ICA+BM

ICA+BM

ICA+BM

ICA+BM

ICA+BM ICA+BM ICA+BM

ICA+BM ICA+BM ICA+BM ICA+BM ICA+BM ICA+BM ICA+BM ICA+BM

ICA+BMICA+BMICA+BMICA+BMICA+BMICA+BMICA+BMICA+BM ICA+BMICA+BMICA+BMICA+BM ICA+BMICA+BMICA+BMICA+BM

ICA+BM ICA+BM ICA+BM ICA+BM ICA+BM ICA+BM ICA+BM ICA+BMICA+BM ICA+BM ICA+BM ICA+BM

ICA+BM ICA+BM ICA+BM ICA+BMICA+BM ICA+BM ICA+BM ICA+BM ICA+BM ICA+BM ICA+BM ICA+BMICA+BM ICA+BM ICA+BM ICA+BMICA+BM ICA+BM ICA+BM ICA+BM


Mask mergingIf the signals in the time domain are correlated, their corresponding masks are merged.

The resulting signal from the merged mask is of higher quality.

Results• Evaluation on real stereo music

recordings, with the stereo recording of each instrument available, before mixing.

• We find the correlation between the obtained sources and the by the ideal binary mask obtained sources.

• Other segregated music examples are available online.


Bas

s

Bas

s D

rum

Gui

tar d

Gui

tar f

Sna

re D

rum

Output1

72% 92%

3% 1% 17%

Output2 5% 1%

55%

4% 14%

Output3 9% 4% 9%

72% 21%

Remaining

14% 3%

32% 23% 48%

% of power 46% 27% 1% 7% 7%

Results

• The segregated outputs are dominated by individual instruments

• Some instruments cannot be segregated by this method, because they are not spatially different.


Conclusion and future work• We have presented an unsupervised method for

segregation of single instruments or vocal sound from stereo music.

• Our method is based on combining ICA and T-F masking.

• The segregated signals are maintained in stereo.• Only spatially different signals can be segregated

from each other. • The proposed framework may be improved by

combining the method with single channel separation methods.


blues from music · microsoft powerpoint - blues from music.ppt author: str_msp created date:...

Documents